Prior to the recent application of stable isotope based GC/MS methodology, little was known about human essential fatty acid metabolism in vivo. Our studies have focused on the metabolic capacities of infants in the first week of life and also on that of human adults. The first phase of this work defined the conversion of linoleic acid to arachidonate and also the conversion of alpha-linolenate to docosahexaenoate in infants of varying gestational ages. The somewhat surprising results were that nearly every infant was capable of both n-3 and n-6 fatty acid interconversions in vivo. Moreover, there was an inverse relationship of gestational age with plasma deuterium enrichment of DHA, in particular; i.e., the least developed infants had the greatest metabolic capability in this respect. This is consistent with the brain growth spurt that occurs in human fetuses during the last trimester. Infants who were small for gestational age had a somewhat diminished metabolic capacity for fatty acids. In this reporting period, a novel multiple-isotope technique that we have termed MultiplE Simultaneous Stable Isotopes, or MESSI, has undergone further development and application. This technique was invented to address the difficult problem of determining the relative efficacy of metabolism of various substrates along a pathway of fatty acid metabolism involving multiple steps. An old and intractable problem has been the direct comparison of metabolism, for example, of linoleate vs. that of gamma-linolenate vs dihommo-gamma-linolenate to form arachidonate. Using the in vivo stable isotope approach and employing NCI GC/MS, one can simultaneously perform the analysis of various isotopomers of arachidonate from multiple precursors providing that suitable isotopes are selected to give a significant mass difference, eg, 5 daltons or more. In the present experiments, rats were given an oral dose of oil containing the following isotopes: 13-C-U-18:2n6, D5-20:3n6, D5-18:3n3, 13-C-U-20:5n3. It was demonstrated that both n-6 fatty acid isotopes were converted to 20:4n6 and that they could be simultaneously measured. In the same animal, the n-3 pathway could also be assessed, both with respect to the 18-carbon and 20-carbon precursor conversions to 22:5n3 and 22:6n3. Thus, the need for four or more separate groups of animals are obviated by this approach with better control since the conditions in separate animals can never be as similar as two comparisons within the same animal at the same time. In connection with these studies, it was important to determine whether either of the stable isotopes led to an decreased rate of metabolism relative to the endogenous compounds. No isotope effect could be detected with deuterated or 13-C labeled linoleic or alpha-linolenic acids. This was the first such in vivo study. Moreover, this approach has now been directly applied to the study of the essential fatty acid metabolism of 18- vs. 20- carbon fatty acids in human infants. Both the NIAAA IRB and the FDA have now approved the use of these multiple stable isotopes in human infants and an initial study of a group of 12 infants has been successfully completed. Physiologic compartmental models were constructed to compare the biosynthesis of 22:6n3 from 18:3n3 and 20:5n3 in plasma. Term neonates were administrated an oral dose of 20 mg of 2H5-18:3n3 and 2 mg of 13C-20:5n3 per kg of BW. Blood was then sampled at 0, 4, 8, 24, 48, 96, 168 hr after administration. 2H5-18:3n3 and 13C-U-20:5n3 as well as their metabolites were simultaneously detected in plasma. A greater rate constant coefficient for the conversion of 2H5-22:5n3 to 2H5-22:6n3 (0.05 hr-1) than for 13C20-22:5n3 to 13C20-22:6n3 (0.014 hr-1) was determined from the model calculations on seven infants. This resulted in an hourly synthetic rate of 47 nmol for the 18:3n3-derived 22:6n3 compared to 17 nmol for the 20:5n3-derived 22:6n3 (P=0.04)). Compartmental modeling is a useful tool for calculating biosynthetic rate parameters that are needed for determining n-3 fatty acid substrate utilization for 22:6n3 supply. A second closely related research project concerns the origins of nervous system DHA. Possible sources are from dietary preformed DHA, from metabolism of the precursor, LNA, or from body stores of DHA. A novel technique has been developed that allows for the quantitative assessment of the amount of DHA accreted from LNA metabolism under various dietary conditions. For this study, it is necessary to control the diet from near birth up to a period where significant brain development has occurred. This has been accomplished thru the use of hand feeding techniques that may be combined with our newly developed artificial feeding approach. An artificial rat milk was developed that was nearly devoid of n-3 fatty acids. The n-3 fatty acids are then added as deuterated-LNA and containing varying levels of DHA. In one major experiment, rat pups were fed diets with 0 or 2% DHA between days 8-29 of life. During this period, it could be calculated that 40% of the newly formed brain DHA in the animals fed D5-LNA as their only source of n-3 fatty acids were derived from preformed DHA and not from LNA metabolism. This was surprising as there was no DHA in the diet; thus, all preformed DHA deposited in the brain must have been derived from other organs via the blood stream. When DHA was added to the diet, there was a pronounced decrease in the rate of LNA metabolism to DHA, a type of end-product inhibition. There was also a higher level of brain DHA in the rats given preformed DHA indicating that metabolism could not provide an adequate source of brain DHA. An attempt was made to determine what the underlying mechanisms for DHA transport into brain and other organs. Lipoproteins were purified and labeled with radiotracers and modified with a tracer levels of phospholipids acylated with DHA, AA or oleic acid (OA). The modified lipoproteins were intravenously injected in mice. The plasma and tissue distribution of the radiotracers were investigated as a function of time and the lipoproteins composition. We found that higher proportion of DHA in LDL results in an enhanced uptake of these lipoproteins by brain and heart. A similar enrichment of LDL in AA or OA did not result in any changes compared to control unaltered LDL. Tissue uptake of HDL did not depend on its fatty acid composition. We next compared the distribution in plasma pools and tissue uptake of 14C-DHA and 3H-(OA) intravenously injected in mice. We found that DHA is rapidly taken up by liver, selectively acylated into triglycerides and released back into the circulation in VLDL. Most of the DHA from VLDL and LDL appeared to be rapidly taken up by extrahepatic organs. This pattern seems to be unique for DHA, because no significant amount of non-essential oleic acid, traced in a similar way, was found in TG and VLDL fractions. In summary, these results point to the important role of VLDL and LDL in transport of DHA to extrahepatic tissues, and to the involvement of liver in the initial selectivity for DHA transport. A novel application of PET imaging for the study of C11-DHA incorporation into brain has been initiated. Brain and heart images from eight individuals have now been obtained and data anlysis has begun. Extensive characterization of the fatty acid input function in plasma has been made in real time for the 11-C-DHA.